{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NB4NLEJ7PIZINSZ66DYFIFY7MR","short_pith_number":"pith:NB4NLEJ7","schema_version":"1.0","canonical_sha256":"6878d5913f7a3286cb3ef0f054171f646f42292bdc0c3db808fc5e947ebbc6ef","source":{"kind":"arxiv","id":"2606.00147","version":1},"attestation_state":"computed","paper":{"title":"RAFT: Data Refinement and Adaptive Distillation for Domain Fine-Tuning with Alleviated Forgetting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hua Zhou, Longbin Yu, Qian Kou, Xiaofeng Shi, Yuduo Li","submitted_at":"2026-05-29T03:06:14Z","abstract_excerpt":"Domain-specific supervised fine-tuning (SFT) often improves in-domain performance at the cost of degrading a model's general capabilities. We view this degradation through two practical gaps in domain SFT: a supervision-compatibility gap, where domain targets differ in style and reasoning format from the original model's natural responses, and a trajectory-preservation gap, where teacher-forced SFT optimizes fixed target tokens without constraining the model's behavior on its own generated prefixes. This process fails to preserve the model's original behavior. We propose RAFT (Data Refinement "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.00147","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-29T03:06:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"86fbc3b673a191e5ef9040d47cf4dd693be7a1ebfbba7a1e579611de4607037d","abstract_canon_sha256":"fdf5c3cbcb99ed907034af8141af2a4b6c0dfd07562c52961f3b132cbfa1d4fb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:19.784726Z","signature_b64":"v1mSQdm4N8ssdXGwoxWLNrenVWkRiOqoAIOkQczqER7LkxHxJHyjTNgadKGm/kG0n/wNUFc/J9DEZvJ+jNxSDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6878d5913f7a3286cb3ef0f054171f646f42292bdc0c3db808fc5e947ebbc6ef","last_reissued_at":"2026-06-02T01:03:19.784322Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:19.784322Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RAFT: Data Refinement and Adaptive Distillation for Domain Fine-Tuning with Alleviated Forgetting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hua Zhou, Longbin Yu, Qian Kou, Xiaofeng Shi, Yuduo Li","submitted_at":"2026-05-29T03:06:14Z","abstract_excerpt":"Domain-specific supervised fine-tuning (SFT) often improves in-domain performance at the cost of degrading a model's general capabilities. We view this degradation through two practical gaps in domain SFT: a supervision-compatibility gap, where domain targets differ in style and reasoning format from the original model's natural responses, and a trajectory-preservation gap, where teacher-forced SFT optimizes fixed target tokens without constraining the model's behavior on its own generated prefixes. This process fails to preserve the model's original behavior. We propose RAFT (Data Refinement "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00147","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.00147/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.00147","created_at":"2026-06-02T01:03:19.784380+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00147v1","created_at":"2026-06-02T01:03:19.784380+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00147","created_at":"2026-06-02T01:03:19.784380+00:00"},{"alias_kind":"pith_short_12","alias_value":"NB4NLEJ7PIZI","created_at":"2026-06-02T01:03:19.784380+00:00"},{"alias_kind":"pith_short_16","alias_value":"NB4NLEJ7PIZINSZ6","created_at":"2026-06-02T01:03:19.784380+00:00"},{"alias_kind":"pith_short_8","alias_value":"NB4NLEJ7","created_at":"2026-06-02T01:03:19.784380+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR","json":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR.json","graph_json":"https://pith.science/api/pith-number/NB4NLEJ7PIZINSZ66DYFIFY7MR/graph.json","events_json":"https://pith.science/api/pith-number/NB4NLEJ7PIZINSZ66DYFIFY7MR/events.json","paper":"https://pith.science/paper/NB4NLEJ7"},"agent_actions":{"view_html":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR","download_json":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR.json","view_paper":"https://pith.science/paper/NB4NLEJ7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00147&json=true","fetch_graph":"https://pith.science/api/pith-number/NB4NLEJ7PIZINSZ66DYFIFY7MR/graph.json","fetch_events":"https://pith.science/api/pith-number/NB4NLEJ7PIZINSZ66DYFIFY7MR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR/action/storage_attestation","attest_author":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR/action/author_attestation","sign_citation":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR/action/citation_signature","submit_replication":"https://pith.science/pith/NB4NLEJ7PIZINSZ66DYFIFY7MR/action/replication_record"}},"created_at":"2026-06-02T01:03:19.784380+00:00","updated_at":"2026-06-02T01:03:19.784380+00:00"}